trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "global_step": 2048,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.24,
      "learning_rate": 0.0008600012539806909,
      "loss": 6.1245,
      "step": 500
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0005184036114706795,
      "loss": 5.0876,
      "step": 1000
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00016650003884818133,
      "loss": 4.9188,
      "step": 1500
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3547716606548965e-06,
      "loss": 4.8719,
      "step": 2000
    }
  ],
  "max_steps": 2048,
  "num_train_epochs": 9223372036854775807,
  "total_flos": 1.1192238096777216e+16,
  "trial_name": null,
  "trial_params": null
}